I would be concerned about NFS locking in failure scenarios
what I do is have an address shared by my syslog machines, and then have them
write the log locally, and relay it to the redundant machine(s) if it arrives on
the shared address.
at a previous job, I use multicast mac for this (CLUSTERIP on linux), at my
current job I have an anycast address shared between 2 machines in each of 2
datacenters for my collectors, and a separate anycast address shared globally
for my relay systems.
David Lang
On Fri, 21 Sep 2018, eSX via rsyslog wrote:
Date: Fri, 21 Sep 2018 00:32:30 +0800
From: eSX via rsyslog <rsyslog@lists.adiscon.com>
To: rsyslog-users <rsyslog@lists.adiscon.com>
Cc: eSX <nay...@gmail.com>
Subject: Re: [rsyslog] rsyslog data on NFS storage and failover
Hi Andriy
I don’t think this solution is very well.
All data saved in NFS means no matter how many rsyslog instance it always
useless because of a single NFS server. Although, multiple rsyslog can keep
rsyslog server more available.
I think you can keep more instance for data storage but not only one. The
failover switch between multiple storage is better if the data is very
important.
Of course, multiple rsyslog is necessary too.
A immature solution can be:
RsyslogA write logs to NFS1, if failed, it relay to rsyslogB, B write to
NFS2. If NFS2 goes down too, just write to local disk for temporary
failover. And you can recover it after NFS back again.
In this case, we have 3 failover rooms, the possibility of three servers
crashing at the same time is very low.
But I also think why NFS? NFS client could block if using an inaccurate
configuration, but not a failure immediately.
Why not relay logs to storage servers directly? Rsyslog has several methods
to do this.
Andriy Tovstik via rsyslog <rsyslog@lists.adiscon.com>于2018年9月20日 周四23:50写道：
I try to setup failover rsyslog configuration with primary and failover
instances for a lot of services. Both rsyslog instances write files to NFS
share to the same location throw omfile module. Looks like solution works
fine - when the primary server fails all messages successfully data is
successfully stored on NFS through the secondary rsyslog server.
But this configuration has a SPOF - NFS server. Is there any ability to
avoid or minimize log messages loss in case of NFS goes down? For example -
any failover local buffer which will store logs until NFS up.
---
Regards, Andrey
_______________________________________________
rsyslog mailing list
http://lists.adiscon.net/mailman/listinfo/rsyslog
http://www.rsyslog.com/professional-services/
What's up with rsyslog? Follow https://twitter.com/rgerhards
NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad
of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you
DON'T LIKE THAT.
--
eSX
_______________________________________________
rsyslog mailing list
http://lists.adiscon.net/mailman/listinfo/rsyslog
http://www.rsyslog.com/professional-services/
What's up with rsyslog? Follow https://twitter.com/rgerhards
NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of
sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T LIKE
THAT.
_______________________________________________
rsyslog mailing list
http://lists.adiscon.net/mailman/listinfo/rsyslog
http://www.rsyslog.com/professional-services/
What's up with rsyslog? Follow https://twitter.com/rgerhards
NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of
sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T LIKE
THAT.
